MDO 3000 Spectrum Analyzer Post Processing?

[ftwang]

I'm wondering if anyone can shed some light regarding the advanced RF analysis of the MDO3000 spectrum analyzer.

From the datasheet, page 9, it seems to suggest that the MDO3000 can save the baseband data.
http://www.tek.com/sites/tek.com/files/media/media/resources/MDO3000-Oscilloscope-Datasheet-1.pdf

Advanced RF analysis
The MDO3000 can save the baseband I and Q data from spectrum analyzer acquisitions to a .TIQ file. These files can then be imported into Tektronix SignalVu-PC software for general purpose modulation and pulse analysis or RSAVu for analysis of commercial wireless standards.

Has anyone worked with this file format, TIQ?  Does that mean that the spectrum analyzer can capture all the IQ baseband data and use software to decode these packets?  Since the MDO3000 has a huge bandwidth, would it be able to capture the the baseband data of multiple protocols simultaneously?  Or is it only able to show the IQ constellation diagram and EVM?

Thanks!

[w2aew]

I'm wondering if anyone can shed some light regarding the advanced RF analysis of the MDO3000 spectrum analyzer.

From the datasheet, page 9, it seems to suggest that the MDO3000 can save the baseband data.
http://www.tek.com/sites/tek.com/files/media/media/resources/MDO3000-Oscilloscope-Datasheet-1.pdf

Advanced RF analysis
The MDO3000 can save the baseband I and Q data from spectrum analyzer acquisitions to a .TIQ file. These files can then be imported into Tektronix SignalVu-PC software for general purpose modulation and pulse analysis or RSAVu for analysis of commercial wireless standards.

Has anyone worked with this file format, TIQ?  Does that mean that the spectrum analyzer can capture all the IQ baseband data and use software to decode these packets?  Since the MDO3000 has a huge bandwidth, would it be able to capture the the baseband data of multiple protocols simultaneously?  Or is it only able to show the IQ constellation diagram and EVM?

Thanks!

Yes, the MDO can save baseband I&Q data.  The MDO converts the incoming RF signal to baseband IQ, centered around the CF of the spectrum analyzer.  The "length" of the IQ data vector is inversely proportional the RBW setting on the spectrum analyzer.  This TIQ file can be fully analyzed in the Tektronix SignalVu-PC vector signal analysis software.  This software can make multiple measurements (spectrums vs time, amplitude/frequency/phase vs. time, analog and digital demodulation including constellation, evm, etc. for many signal types). 

The MDO4000 can capture longer RF acquisitions vs. time - much longer than the spectrum-time defined by the RBW.

[ftwang]

Thanks Alan!  I'd be interested if you can add this feature to your review of the MDO3000 series if you have time.  Btw, love your youtube channel!

[w2aew]

Thanks Alan!  I'd be interested if you can add this feature to your review of the MDO3000 series if you have time.  Btw, love your youtube channel!

Thanks!

By the way, I did include that in my quick intro to the MDO4000B "improvements" video.  On the MDO4000B, SignalVu-PC can connect directly to scope and transfer the IQ data via a "live-link".  This feature isn't available for the MDO3000 - but you can definitely save a TIQ file, then read it into SignalVu-PC to accomplish the same thing.

[majbthrd]

On the MDO4000B, SignalVu-PC can connect directly to scope and transfer the IQ data via a "live-link".  This feature isn't available for the MDO3000 - but you can definitely save a TIQ file, then read it into SignalVu-PC to accomplish the same thing.

Alan, I'm inferring that the procedure for saving to a TIQ file is the one shown in your video "Using MDO4000 to capture 802.11 traffic and export for analysis using RSAVu".  (One triggers a capture where the RF functionality is active and then saves a baseband waveform to USB... presumably of the entire "RF Acquisition Time" that has been captured.)

I don't have a MDO3000, but am puzzled as to what the equivalent procedure would be for that unit... and I've looked at every piece of documentation I can find.

Based on the manual and SCPI reference, it seems like the only RF mode for the MDO3000 is the spectrogram display.  It is not apparent how one obtains a TIQ file from that.  I presume each "slice" of the spectrum is only a relatively tiny number of data points (the "spectrum time"?).  The aggregate number of slices (198 for the MDO3000; 328 or 148 for the MDO4000) would be much more data, but would presumably be from discontinuous periods of time.

Also, pushing my luck further still... there is a helpful app note "Fundamentals of the MDO4000B Series Mixed Domain Oscilloscope" that indicates the "RF Acquisition Time" for the MDO4000B.  (I presume the MDO4000 figures are half those?)  Would these time figures represent the length of the TIQ file?  Are there any equivalent figures for the MDO3000?

I also notice that there is documentation for a MATLAB compatible baseband file format (MDO4000B specific?); perhaps this is the format to save in if one wants to post process the data as suggested by the original poster?

Thanks in advance for any insight you can share and thanks for posting the videos.  It is genuinely helpful to see the instruments in operation.

[w2aew]

On the MDO4000B, SignalVu-PC can connect directly to scope and transfer the IQ data via a "live-link".  This feature isn't available for the MDO3000 - but you can definitely save a TIQ file, then read it into SignalVu-PC to accomplish the same thing.

Alan, I'm inferring that the procedure for saving to a TIQ file is the one shown in your video "Using MDO4000 to capture 802.11 traffic and export for analysis using RSAVu".  (One triggers a capture where the RF functionality is active and then saves a baseband waveform to USB... presumably of the entire "RF Acquisition Time" that has been captured.)

I don't have a MDO3000, but am puzzled as to what the equivalent procedure would be for that unit... and I've looked at every piece of documentation I can find.

Based on the manual and SCPI reference, it seems like the only RF mode for the MDO3000 is the spectrogram display.  It is not apparent how one obtains a TIQ file from that.  I presume each "slice" of the spectrum is only a relatively tiny number of data points (the "spectrum time"?).  The aggregate number of slices (198 for the MDO3000; 328 or 148 for the MDO4000) would be much more data, but would presumably be from discontinuous periods of time.

Also, pushing my luck further still... there is a helpful app note "Fundamentals of the MDO4000B Series Mixed Domain Oscilloscope" that indicates the "RF Acquisition Time" for the MDO4000B.  (I presume the MDO4000 figures are half those?)  Would these time figures represent the length of the TIQ file?  Are there any equivalent figures for the MDO3000?

I also notice that there is documentation for a MATLAB compatible baseband file format (MDO4000B specific?); perhaps this is the format to save in if one wants to post process the data as suggested by the original poster?

Thanks in advance for any insight you can share and thanks for posting the videos.  It is genuinely helpful to see the instruments in operation.

That video that uses RSAVu is pretty dated.  The RSAVu tool is really not the best tool to use for this.  SignalVu-PC is the replacement, many more features and a really nice user interface.  The "process" is the same though.

As mentioned above, the MDO3000 does not give you direct control over the RF acquisition length, since there are no RF vs. Time traces like there are on the MDO4000.  On the MDO3000, the RF acquisition length is determined by the RBW and window selection.  For the default Kaiser window, the RF vector length is 2.23/RBW.  To save the IQ data, you'd hit the Save Menu key, then the Save Waveform bezel key.  They you can select the Baseband I&Q as the "waveform" to save.  Under File Details, you can select the TIQ format or a MATLAB .mat file format.  The TIQ can be read by  SignalVu-PC directly.

You've got the details right on the MDO4000 and 4000B.  The MDO3000 doesn't give you control over the RF acquisition length like the 4000 does.  On the 3000, you only get "one spectrum's" worth of IQ.  Thus, making the RBW narrow is the only way to increase the RF acquisition length.  The 3000 allows you to set the RBW as low as 1/200,000th of the SPAN.  It results in about a half million IQ samples if I recall correctly.  Once in SignalVu-PC, you can choose a different RBW to build a spectrogram, as well as create amplitude, frequency, and phase vs. time traces, and various modulation analysis, etc.  It is important to note that the IQ data from the 3000 does not include the amplitude and phase vector calibration and correction that the 4000B has, so wideband modulation EVM won't be as good as the 4000B.

Hope this helps.

FYI - here's an example using the MDO4000B, with a "live-link" of baseband IQ data into SignalVu-PC for vector signal analysis, including WLAN demodulation:

[majbthrd]

Hope this helps.

Thanks much, that definitely helps.

Yes, I recognize the RSAVu user interface as similar to, if not the same as, the Tek RSA that I had access to a few years ago.  The SignalVu interface looks much nicer and seems to have more features.

500k complex samples with a span of 125MHz implies a best case duration of 4ms on the MDO3000 (and proportionally less at higher spans).

That isn't a ton of time, and I'd much prefer the 158ms of the MDO4000B (and calibration/correction to boot), but 4ms might be just enough to catch a signal transition of interest.  What is missing to make it really useful, though, is a trigger mechanism.  This seems a non-starter since there appears to be no way to trigger the RF acquisition (alas, not even through the optional digital pod).

I can see why Tek wouldn't want to risk cannibalizing MDO4000 sales by putting too much in the MDO3000, but perhaps we can hope the MDO3000 "B" might have a trigger option (such as RF power, digital pod, or protocol trigger on digital pod).

edit to add: Even if without a trigger (which would definitely be desirable), I think there is opportunity to milk the IQ data for more than just spectrum/spectrogram information.  With a spectrum analyzer, one just sees a peak, and what that peak is can be a mystery.  However, that same digital information can also be mined to reveal the type of modulation, the constellation size, etc., etc.  It may not be measurement-grade, but it tells you more about the signal behavior.

[w2aew]

Hope this helps.

Thanks much, that definitely helps.

Yes, I recognize the RSAVu user interface as similar to, if not the same as, the Tek RSA that I had access to a few years ago.  The SignalVu interface looks much nicer and seems to have more features.

500k complex samples with a span of 125MHz implies a best case duration of 4ms on the MDO3000 (and proportionally less at higher spans).

That isn't a ton of time, and I'd much prefer the 158ms of the MDO4000B (and calibration/correction to boot), but 4ms might be just enough to catch a signal transition of interest.  What is missing to make it really useful, though, is a trigger mechanism.  This seems a non-starter since there appears to be no way to trigger the RF acquisition (alas, not even through the optional digital pod).

I can see why Tek wouldn't want to risk cannibalizing MDO4000 sales by putting too much in the MDO3000, but perhaps we can hope the MDO3000 "B" might have a trigger option (such as RF power, digital pod, or protocol trigger on digital pod).

edit to add: Even if without a trigger (which would definitely be desirable), I think there is opportunity to milk the IQ data for more than just spectrum/spectrogram information.  With a spectrum analyzer, one just sees a peak, and what that peak is can be a mystery.  However, that same digital information can also be mined to reveal the type of modulation, the constellation size, etc., etc.  It may not be measurement-grade, but it tells you more about the signal behavior.

Yes, the RSAVu interface was the user interface on the older RSA3000 series of realtime spectrum analyzers.  These basically went away 4-5 years ago or so.  SignalVu-PC is the same user interface that is on the latest generation RSA5000 and 6000 series realtime spectrum and signal analyzers, and is also available to run directly on the performance scopes (that run windows), the DPO/MSO 7000 and 70000 series - ranging from 500MHz to 33GHz analog BW.

I've been pestering the product line to look into triggering the rf capture on the MDO3000, but I don't know if it can/will happen. 

You can do more than 4ms on the MDO3k.   I just did a 1MHz span, 30Hz RBW capture resulting in 74ms long IQ record.

[majbthrd]

I've been pestering the product line to look into triggering the rf capture on the MDO3000, but I don't know if it can/will happen. 

You can do more than 4ms on the MDO3k.   I just did a 1MHz span, 30Hz RBW capture resulting in 74ms long IQ record.

Thanks for trying that.  That sounds very promising.

[RJdaMoD]

Hi,

sorry to hijack this thread, but i'm currently checking the specs of the mdo and was wondering if one could also export the baseband data (or the equivalent sampling data of the rf) via scpi or other intefaces.
How does SignalVu this?

Thanks and regards
RJ

[majbthrd]

i'm currently checking the specs of the mdo and was wondering if one could also export the baseband data (or the equivalent sampling data of the rf) via scpi or other intefaces.

I was interested in that too and have some information to share; I too found the Tek documentation lacking and had to experiment on real gear.

Your mileage may vary...

At least for the MDO4000B (v3.14 firmware), there was an undocumented baseband data download function via its built-in web page.  There is no mention of it in any Tek documentation that I could find, but it is clear as day in one of the web page options.

However, I have not yet been able to get seat time with either the MDO3000 or the older MDO4000 to confirm if it has the same functionality.

If memory serves, retrieving data programmatically from the web interface was a bit problematic because of some HTTP argument passing, but it seemed feasible.  (It was more complicated than a wget, though.)

As best as I can tell, the internal web page obtains its data from an undocumented SCPI command ":RF:RF_BB_IQ?".  Issuing the command caused a TIQ file to be returned.  It is a large amount of data and it arrives in spurts, so it may break some SCPI client implementations.  (I think the MATLAB format is probably also possible, but I didn't have the time to figure that out.)  Again, I have only had the opportunity to try this on a MDO4000B.

I guess the risk here is that by listing that undocumented SCPI command on a forum, Tek might remove it, but my hope is that Tek realizes the value in having such a command available to users.  There are things like Automated Test and research where it is desirable to be able to pull such data without having to save everything to USB thumb drives first.

[RJdaMoD]

Thanks for the quick reply, that sounds promising.
I wonder why they do not publish this, as for example rhode&schwarz do this for their spectrum analyzers.

[Omicron]

With the base version of SignalVu-PC now free I thought I'd have a play around with it. I'm sure it's a joy to use with the live link devices, but I can't figure out how to use it with TIQ files from an MDO4000 (non B). I was trying to get a time view, spectrum view and spectrogram with linked navigation between the 3. I had it working at some point, but hell if I know how. It stopped working equally mysteriously. The manual isn't of much help as it seems to focus on the live link stuff. I'd sure welcome a tutorial!

Edit: It's starting to make more sense now. The magic I was missing is all in the "replay" buttons it seems :-)

[luisprata]

The "Fundamentals of the MDO4000B Series Mixed Domain Oscilloscope" makes me think there is two memory spaces, before and after DDC.
I wonder if there is a way to get this memory from MDO3000.